This is a case-control epidemiologic study of brain tumors occurring in children. Cases are those children (ages 1 day to 16 years) with a brain tumor who reside in the catchment area of eight "cancer" registries included in the S.E.E.R. program. Three controls will be selected for each case by random digit dialing, matching for age, (plus or minus 3 years), sex and race. A one and one-half hour long questionnaire will be administered to both parents by an interviewer in the parents' home. The questionnaire includes requests for information about demographic features, parents' exposure to possible teratogens and mutagens, familial history of tumor and birth defects, and for approximately 400 cases and 1,200 controls. This sample size is expected to be sufficient for the study of two of the most common (groups of) tumors (cerebellar gliomas and medulloblastomas).